The efficacy of the triple airway manoeuvre (mouth opening, head extension and jaw thrust) for insertion of the laryngeal mask airway (LMA) was compared with the standard insertion method. One hundred paralyzed patients were allocated randomly into two groups: in the control group (n = 50) the LMA was inserted by the standard method and in the other (TAM group, n = 50) by the triple airway manoeuvre. In ten patients of each group, the position of the LMA and the epiglottis was assessed radiographically before insertion, after insertion but before cuff inflation, and after cuff inflation. In all patients the position war examined using fibrescopy before and after cuff inflation. The mean distance between the epiglottis and the posterior pharyngeal wall, measured radiographically before LMA insertion, was greater in the TAM group (16.3 (SD 4.3) mm) than in the central group (7.0 (2.8) mm) (P < 0.001). Before cuff inflation, radiography and fibrescopy showed that the LMA compressed the epiglottis downwards more frequently in the control group. After cuff inflation the glottis was completely visible fibreoptically in 66% in the TAM group, compared with 14% in the control group (P < 0.001). Complete downfolding of the epiglottis was seen in 10% in the control group and none in the TAM group (P < 0.05). We conclude that in paralyzed patients LMA insertion with the triple airway manoeuvre provides wider pharyngeal space and decreases the incidence of epiglottic downfolding by the LMA compared with the standard method.